By Michael Domanic, Section Head of AI

At the end of last year, I published five predictions for how AI would reshape business in 2026. One of them: by the end of 2026, it will be table stakes to have a Head of AI in your business.

I'll take the win. IBM surveyed 2,000 CEOs across 33 countries and found the share of organizations with a Chief AI Officer jumped from 26% to 76% in a single year. Job postings for the role are up 340% since 2023. The debate about whether this role should exist is over.

Here's the problem: Most of the companies are hiring the wrong person.

The job description trap

I see a lot of Head of AI job descriptions, because companies will often send them to me and ask what I think. 

Most are wrong in the same way: They’re too technical. 

They ask for PhDs in data science or computer science, 20+ years leading IT organizations, or experience with the AI features inside function-specific point solutions.

You can see the reasoning: AI is technology, so hire a technologist. And those aren't bad skills, but they map to someone who can run a technology deployment and that’s not the job. The Head of AI job is building and leading a scalable, durable program around the largest workplace transformation in business history.

We've run this experiment before - during the Chief Digital Officer wave of the early 2010s. That role saw rapid adoption, unclear mandates, external hires who never understood the business, and widespread reassignment or quiet departure within 18 to 24 months. 

The role eventually shifted away from figurehead evangelists toward people who drove real implementation. The same thing is about to happen now - a market full of expensive AI executives who evangelized for a year and got reassigned. If you want to avoid making that hire, you need to know what the right one actually looks like.

The four non-negotiables

When companies ask me what to actually screen for (I get this question all the time), I give them four key characteristics:

1. Deep fluency and proficiency in AI. This is the obvious one, and most companies already recognize it. You want someone operating at or very near the frontier. They're automating workflows and building complicated agents in their personal and professional life, and they spent the weekend turning three Mac Minis into a tiny open-weight model cluster (because that’s what weekend fun looks like, duh?) 

2. Strong understanding of the business. They know enough to be dangerous in every corner of your company. Maybe they've never worked in marketing, but they understand what marketing does and how each part of that team contributes to the business. Same for HR, finance, ops, everywhere else. These people see the whole business and how the parts come together. And it's a real advantage to hire someone who knows your type of business specifically. I've worked in enterprise SaaS and consulting for a long time, and I know those models cold. I would bring far less value as the Head of AI at a consumer goods company, because I don't know that business all that well. Find the person who knows yours.

3. Systems thinking that gets complicated things done. Companies too often file this role under technology, when it belongs under change management. An effective Head of AI knows how to work relationships, how to hold people accountable for deliverables, when to exercise ownership, and when to delegate. 

4. Creativity. This is the least intuitive and by far the most important. Everyone I know who’s effective in this job is deeply creative. They can take the abstract nature of generative AI and turn it into meaningful value across the business. They operate in a true zero-to-one environment, building roadmaps that don't exist in any playbook. This is the hardest to interview for, so ask open-ended questions about problems you haven’t solved yet, and see their brain work. 

If you ignore the last three characteristics, you risk hiring a technology leader for a role that is really about business fluency, organizational change, and invention. This usually results in running the deployment backwards: procure a platform, roll it out, train everyone, then go hunting for ROI to justify what was already bought. And that will fail. 

The best candidate is probably already in your building

None of these four characteristics shows up on a résumé, which is why the résumé is the wrong place to look. 

Look at behavior instead. In most companies, this person already exists somewhere in the org, and you can identify them because they're already transforming the work of the people around them without being asked.

It doesn’t matter what department they’re in. Demonstrated building, real business fluency, and interpersonal influence will beat an impressive external résumé that has never operated inside your business. And to be clear, none of this is a rule against technical candidates. Some of the best people doing this job have deeply technical backgrounds - they just also happen to be creative, business-minded systems thinkers. Look across marketing, finance, engineering, product, and ops for candidates - they’re the ones posting in your AI Slack channel to help others along. 

And if you're evaluating candidates right now, internal or external, run this screen. Ask them to show you something they built with their own hands. Ask them to pitch the business case to your least technical executive. Then ask the people they last worked with what changed because they were there. Most candidates can pass one of these questions, but the job is all three. 

See you next week,

Michael

Your fellow Head of AI

‍